Skip to main content

Full text of "Race Mixture in the Roman Empire"

See other formats


Early Journal Content on JSTOR, Free to Anyone in the World 

This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in 
the world by JSTOR. 

Known as the Early Journal Content, this set of works include research articles, news, letters, and other 
writings published in more than 200 of the oldest leading academic journals. The works date from the 
mid-seventeenth to the early twentieth centuries. 

We encourage people to read and share the Early Journal Content openly and to tell others that this 
resource exists. People may post this content online or redistribute in any way for non-commercial 

Read more about Early Journal Content at 
journal-content . 

JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people 
discover, use, and build upon a wide range of content through a powerful research and teaching 
platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit 
organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please 

Volume XXI~\ July, 1916 {Number 4 

Mtmm If isitfltial Utvxtw 


THERE is one surprise that the historian usually experiences 
upon his first visit to Rome. It may be at the Galleria Lapi- 
daria of the Vatican or at the Lateran Museum, but, if not else- 
where, it can hardly escape him upon his first walk up the Appian 
Way. As he stops to decipher the names upon the old tombs that 
line the road, hoping to chance upon one familiar to him from his 
Cicero or Livy, he finds praenomen and nomen promising enough, 
but the cognomina all seem awry. L. Lucretius Pamphilus, A. 
Aemilius Alexa, M. Clodius Philostorgus do not smack of freshman 
Latin. And he will not readily find in the Roman writers now ex- 
tant an answer to the questions that these inscriptions invariably 
raise. Do these names imply that the Roman stock was completely 
changed after Cicero's day, and was the satirist recording a fact 
when he wailed that the Tiber had captured the waters of the 
Syrian Orontes? If so, are these foreigners ordinary immigrants, 
or did Rome become a nation of ex-slaves and their offspring? Or 
does the abundance of Greek cognomina mean that, to a certain 
extent, a foreign nomenclature has gained respect, so that a Roman 
dignitary might, so to speak, sign a name like C. Julius Abascantus 
on the hotel register without any misgivings about the accommo- 
dations ? 

Unfortunately, most of the sociological and political data of the 
empire are provided by satirists. When Tacitus informs us that in 
Nero's day a great many of Rome's senators and knights were de- 
scendants of slaves and that the native stock had dwindled to sur- 
prisingly small proportions, we are not sure whether we are not to 
take it as an exaggerated thrust by an indignant Roman of the old 
stock. At any rate, this, like similar remarks equally indirect, re- 
ceives totally different evaluation in the discussion of those who have 
treated of Rome's society, like Friedlander, Dill, Mommsen, Wallon, 

AM. HIST. REV., VOL. XXI. — 45. (689) 

690 Tenney Frank 

and Marquardt. To discover some new light upon these funda- 
mental questions of Roman history, I have tried to gather such frag- 
mentary data as the corpus of inscriptions might afford. This evi- 
dence is never decisive in its purport, and it is always, by the very 
nature of the material, partial in its scope, but at any rate it may 
help us to interpret our literary sources to some extent. It has at 
least convinced me that Juvenal and Tacitus were not exaggerating. 
It is probable that when these men wrote a very small percentage of 
the free plebeians on the streets of Rome could prove unmixed 
Italian descent. By far the larger part — perhaps ninety per cent. — 
had Oriental blood in their veins. 

My first quest was for information about the stock of the ordi- 
nary citizen of Rome during the empire. In the Corpus of Latin 
Inscriptions 1 the editors, after publishing the honorary and sepul- 
chral inscriptions of the nobles and military classes, followed by 
those of the slaves and humble classes which occur in the columbaria, 
gave the rest of the city's sepulchral inscriptions (19,260) in alpha- 
betical order. 2 Of these I read the 13,900 contained in volume VI., 
parts 2 and 3, which, despite the occurrence of some slaves as well 
as of some persons of wealth, represent on the whole the ordinary 
type of urban plebeians. A mere classification of all these names 
into lists of natives on the one hand and slaves and foreigners on 
the other would be of little service, since, obviously, transient for- 
eigners are of little importance in estimating the stock of the per- 
manent population of Rome, and we must face the question at once 
whether or not the slave and freedman stock permanently merged 
into the civil population. Furthermore, such lists will be at every- 
one's hand as soon as the index of the sixth volume of CIL. is pub- 
lished. In reckoning up the foreign stock, therefore, I have counted 
only those who, according to the inscriptions, were presumably born 
at Rome. A somewhat arbitrary definition of limits was necessary 
since we are seldom given definite information about the place of 
birth, but as I have used the same classification for the free-born as 
for the slave-born the results are valid for our purposes. For in- 
stance, in getting statistics of birth, I have included all children 
under ten years of age, assuming that slave children under that age 
would rarely be brought in from abroad ; and if slaves of this class 
are counted, the free-born of the same class must also be reckoned 
with. I have also included slave and free-born children who appear 
to be with father, mother, brother, or sister at Rome, since presum- 
ably they would have been sundered from their family if they had 

1 CIL., vol. VI., parts 2, 3, 4. 

2 Vol. VI., part 4 2 , published in 1902, contains 2572 additional inscriptions of 
this class. 

Race Mixture in the Roman Empire 69 1 

been brought in from the foreign market; and again, in order to 
reach fair results, the corresponding persons of free birth are 
counted. For reasons which will presently appear I have accepted 
the Greek cognomen as a true indication of recent foreign extrac- 
tion, and, since citizens of native stock did not as a rule unite in 
marriage with liberti, a Greek cognomen in a child or one parent is 
sufficient evidence of status. As is well known, certain Latin cog- 
nomina, e.g., Salvius, Hilarus, Fortunatus, were so frequently borne 
by slaves and freedmen that they were apt to be avoided by the 
better classes. Nevertheless, since no definite rule is attainable in 
the matter, I have credited the bearers of all Latin names to the 
native stock in all cases of doubt. 3 

Classifying in this way the names of the aforesaid 13,900 in- 
scriptions of volume VI., parts 2 and 3, we find that of the 4485 
persons apparently born at Rome, 3723 (eighty-three per cent.) fall 
into the list which by our criteria represents foreign extraction. 
This figure is probably not far from correct, but I think it would 
be raised somewhat if it were possible to decide what proportion of 
Latin cognomina conceals slaves and liberti. For instance, a name 
like Q. Manlius Restitutus (VI. 22015) would usually pass with 
little suspicion. But the inscription also names his father, mother, 
wife, and two sons, all of whom have Greek cognomina. Because 
of his parentage I have classed him as of foreign stock, but there are 
scores of brief inscriptions in which the necessary facts are not pro- 
vided. In these the subject had to be classed, however erroneously, 
as Latin. 

In order to reckon if possible the margin of error in cases like 

3 In epigraphical discussions one constantly meets with the statement that 
freedmen were compelled to indicate their status by the designation lib. or /. and 
that therefore the occurrence of the tria nomina without such designation is 
proof of free birth. Unfortunately, this rule, if indeed it was one, was so fre- 
quently broken, that it must be employed with caution. There are hundreds of 
obvious exceptions where tria nomina of respectable appearance impose upon the 
reader until at the end of the inscription the dedicant's designation of patronus 
or contubernalis or conlibertus betrays the real status, e. g., VI. 7849, i455°> 
16203, '7562, 20675, 20682, 22299, 22606, 23927, 23989. Again, numerous bearers 
of faultless tria nomina fall under strong presumption of being freedmen because 
of some official title like sevir or because their sons prove to belong to one of the 
city tribes; cf. X. 690, 4620, 6677; VI. 12431, 14045, 20079. Finally, there are 
many instances like 14018. Here a man gives the name of a large family (all 
with tria nomina) including children and a grandchild, but only the youngest, 
Caesonia M. F. Prima, a child of seven months, bears the F which definitely 
indicates free birth. Apparently the other members of the family were not en- 
titled to the designation. Compare also 20123, 20339, 23813. Since in cases of 
doubt I have been compelled to credit bearers of Latin tria nomina to the native 
stock, it will appear that this group has more than received full credit in the 
accompanying lists. 

692 Tenney Frank 

this, I have attempted to test the respectability of Latin cognomina, 
but with rather unsatisfactory results. I counted all the names of 
slaves and freedmen in the indexes of volumes V., IX., XIV., and 
over a thousand in volume VI., in order to get a group of five thou- 
sand bearing the prevalent slave-names. More than half (2874) 
have Greek names, the most popular of these being Eros (58 times), 
Pamphilus (36), Antiochus (34), Hermes (30), Alexander (28), 
Philomusus (26), Onesimus (22), Philargyrus (21), names, most 
of which were also very popular among free Greeks and Asiatics. 
Two thousand one hundred and twenty-six have Latin names, some 
of which occur with remarkable frequency, e.g., Felix (97) , Hilarus 
-a (64-53), Faustus -a (58-33), Salvius -a (38-18), Fortunatus -a 
(29-15), Primus -a (51-47), Secundus -a (25-34), Tertius -a (18- 
18), Auctus -a (24-15), Vitalis (36), Januarius -a (22-6). Now, 
if we compare these Latin names with those borne by better-class 
Roman plebeians, by the pretorian guards, for instance (though 
many descendants of slaves served even in the pretorian guards), we 
find, despite a certain overlapping, quite a striking difference. Ap- 
parently some names had acquired such sordid associations that they 
were in general avoided by ordinary plebeians. The favorite names 
on the pretorian lists are Maximus, Proculus, Severus, Verus, 
Capito, Justus, Celer, Marcellus, Clemens, Victor, and the like. We 
may not say that any Latin name was confined wholly to slaves, nor 
would it be possible to give any usable list of relative percentages, 
but we may at least say that the Romans recognized such names as 
Salvius, Hilarus, Fortunatus, Optatus, Auctus, Vitalis, Januarius, 
as being peculiarly appropriate to slaves ; and Felix, Faustus, Primus, 
Primitivus, and a few others must have cast some suspicion upon 
the bearer. After reviewing in this light the seventeen per cent, of 
possible claimants of Latin origin in the alphabetical list of inscrip- 
tions in volume VI., parts 2 and 3, I have little doubt that a third 
of these would, with fuller evidence, be shifted into the class of 

On the other hand, the question has been raised whether a man 
with a Greek cognomen must invariably be of foreign stock. Could 
it not be that Greek names became so popular that, like Biblical and 
classical names to-day, they were accepted by Romans of native 
stock ? In the last days of the empire this may have been the case ; 4 

* There are not enough datable inscriptions available to show whether the 
Greek cognomen gained or lost respectability with time. Obviously it may in 
general be assumed that most of the freedmen who bore the gentile name of Aelius 
and Aurelius belong to a later date than the general group of those named Julius 
and Claudius. If we may use this fact as a criterion we may decide that there 
was little difference between the first and the second century in this matter, since 
the proportion of Greek cognomina is about the same in the two groups. 

Race Mixture in the Roman Empire 693 

but the inscriptions prove that the Greek cognomen was not in good 
repute. I have tested this matter by classifying all the instances in 
the 13,900 inscriptions (there are 1347) where the names of both 
father and son appear. 5 From this it appears that fathers with 
Greek names are very prone to give Latin names to their children, 
whereas the reverse is not true. The statistics are as follows : 

Greek cognomen Latin cognomen 

Father 859 488 

Greek Latin Greek Latin 

Son 460 399 53 435 

This means that in one generation Greek names diminish from sixty- 
four per cent, to thirty-eight per cent., or that forty-six per cent, of 
the fathers with Greek names .give their sons Latin names, while 
only eleven per cent, of the Latin fathers give their sons Greek 
names. And this eleven per cent, dwindles upon examination 
into a negligible quantity. For instance, in seventeen of the 
fifty-three cases the mother's name is Greek, which betrays the true 
status of the family; and in ten other instances the son's gentile 
name differs from that of the " father ", who is, therefore, probably 
a stepfather. In almost all of the other twenty-six instances, the 
inscription is too brief to furnish a fair criterion for judging. 
Clearly the Greek name was considered as a sign of dubious origin 
among the Roman plebeians, and the freedman family that rose to 
any social ambitions made short shrift of it. For these reasons, 
therefore, I consider that the presence of a Greek name in the imme- 
diate family is good evidence that the subject of the inscription is 
of servile or foreign stock. The conclusion of our pros and cons 
must be that nearly ninety per cent, of the Roman-born folk repre- 
sented in the above-mentioned sepulchral inscriptions of CIL., vol- 
ume VI., parts 2 and 3, are of foreign extraction. 

Who are these Romans of the new type and whence do they 
come? How many are immigrants, and how many are of servile 
extraction? Of what race are they? Seneca happens to make a 
remark which is often quoted as proof of extensive immigration to 
Rome. He writes to his mother in derision of Rome : 

Of this crowd the greater part have no country; from their own free 
towns and colonies, in a word, from the whole globe, they are congre- 
gated. Some are brought by ambition, some by the call of public duty, 

5 It is difficult to secure usable statistics in the case of women, since their 
cognomina may come from almost any relative or near friend. However, an 
examination of the indexes of names will show that the Greek cognomen was 
relatively no more popular among the women than among the men. 

694 Tenney Frank 

or by reason of some mission, others by luxury which seeks a harbor rich 
and commodious for vices, others by the eager pursuit of liberal studies, 
others by shows, etc. 6 

Seneca apparently refers in large part to visitors, but also to im- 
migrants. In so far as he has transients in mind we are not con- 
cerned with the passage, for such people did little to affect the per- 
manent racial complexion of Rome's civil population. A passage in 
Juvenal's third satire is perhaps more to the point, for he seems to 
imply that the Oriental has come to stay. 

While every land . . . 

daily pours 
Its starving myriads forth. Hither they come 
To batten on the genial soil of Rome, 
Minions, then lords of every princely dome, 
Grammarian, painter, augur, rhetorician, 
Rope-dancer, conjurer, fiddler, and physician. 

This passage clearly suggests that foreigners of their own free will 
have drifted to Rome in great numbers to make it their place of live- 
lihood and their permanent abode. I cannot here treat the whole 
problem, but, while agreeing that the implication of this passage is 
true to a certain degree, I would question whether the generalities 
in it are not too sweeping. It may well be that many of the ex-slave 
rabble who spoke the languages of the East imposed upon the un- 
critical by passing as free-born immigrants. Even freedmen were 
not beyond pretending 7 that they had voluntarily chosen slavery 
as a means of attaining to Roman citizenship by way of the vindicta. 
At any rate, the Roman inscriptions have very few records of free- 
born foreigners. Such men, unless they attained to citizenship, 8 
ought to bear names like that in no. 17171, Dis man. Epaeneti, 
Epaeneti F. Ephesio, but there are not a dozen names of this sort to 
be found among the inscriptions of volume VI., parts 2 and 3. Nor 
need we assume that many persons of this kind are concealed among 
the inscriptions that bear the tria nomina, for immigrants of this 
class did not often perform the services for which the state granted 
citizenship. There could hardly have been an influx of foreign free- 
born laborers at Rome, for Rome was not an industrial city and was 
more than well provided with poor citizens who could not compete 
with slaves and had to live upon the state's bounty. Indeed, an ex- 
amination of the laborious article by Kiihn 9 fails to reveal any free- 

6 Ad Helviam, 6. 
? Petronius, 57. 

8 This criterion fails of course after citizenship was given to the provincials 
in the third century, but when Rome's population was decreasing there probably 
was not a heavy immigration. 

9 De Opificum Romanorum Condicione (1910). 

Race Mixture in the Roman Empire 695 

born foreigners among the skilled laborers of the city. In regard to 
shop-keepers, merchants, and traders we may refer to a careful dis- 
cussion by Parvan. 10 He has convincingly shown that the retail 
trade was carried on at Rome, not by foreigners but by Romans of 
the lower classes, mostly slaves and freedmen, and that while the 
provincials of Asia and Egypt continued throughout the empire to 
carry most of the imports of the East to Rome, the Roman houses 
had charge of the wholesale trade in the city. The free-born for- 
eigner did not make any inroad upon this field. However, in various 
arts and crafts, such as those mentioned by Juvenal, the free immi- 
grant could gain a livelihood at Rome. Some of the teachers of 
rhetoric, philosophy, and mathematics, some of the doctors, sculptors, 
architects, painters, and the like, were citizens of the provincial 
cities who went to Rome for greater remuneration. But even most 
of these professions were in the hands of slaves and freedmen who 
had been given a specialized education by their masters. In volume 
VI., part 2, which contains the sepulchral inscriptions classified ac- 
cording to arts and crafts, there is very little trace of the free-born 
foreigner. Among the fifty inscriptions of medici, for instance, 
only two, 9563, 9597, contain sure instances of such foreigners. 
Among the grammatici, rhetores, argentarii, structores, and pictores, 
where they might well be expected, I find no clear case. It is evi- 
dent then that the sweeping statements of men like Juvenal and 
Seneca should not be made the basis for assuming a considerable 
free-born immigration that permanently altered the citizen-body of 
Rome. These writers apparently did not attempt to discriminate 
between the various classes that were speaking foreign jargons on 
the streets of Rome. As a matter of fact, this foreign-speaking 
population had, for the most part, it seems, learned the languages 
they used within the city itself from slaves and f reedman parents of 
foreign birth. 

If now this great crowd of the city was not of immigrant stock, 
but rather of servile extraction, the family life of the slaves must 
have been far more conducive to the propagation of that stock than 
is usually assumed, and, furthermore, manumission must have been 
practised so liberally that the slave-stock could readily merge into 
the citizen-body. On the latter question our sources are satisfac- 
tory ; on the former, they have little to say. From Varro (II. i. 26 
and x. 6) and Columella (I. 8, 19) it has been well known that 
slaves on farms and pasture-lands were expected to marry and have 
offspring. The Romans considered this good economy, both be- 
cause the stock of slaves increased thereby and because the slaves 

1° Die Nationalitat der Kaufleute im Romischen Kaiserreich (1909). 

696 Tenney Frank 

themselves remained better satisfied with their condition. However, 
partly because there exists no corresponding statement regarding 
slaves in the city, partly because of a reckless remark made by Plu- 
tarch that Cato restricted the cohabitation of his slaves, partly, too, 
because service in the city household is supposed to have been very 
exacting, the prevalent opinion seems to be that the marriage of 
slaves in the urban familia was unusual. Hence the statement is 
frequently made that slavery died perforce when the pax Romana 
of the empire put an end to capture by warfare. 

Fortunately the columbaria of several Roman households provide 
a fairly reliable record regarding the prevalence of marriage among 
city slaves. In CIL., VI. 2, some 4500 brief inscriptions are given, 
mainly from the rude funeral urns of slaves and poor freedmen of 
the first century of the empire. About one-third of these are from 
the columbaria of the Livii, Drusi, Marcelli, Statilii, and Volusii, 
aristocratic households where, presumably, service would be as ex- 
acting as anywhere, discipline as strict, and concern for profits from 
the birth of vernae as inconsiderable as anywhere. Furthermore, 
these inscriptions date from a time when slaves were plentiful and 
the dearth of captives generally assumed for a later day cannot be 
posited. Nevertheless, I believe that anyone, who will studiously 
compare the record of offspring in this group of inscriptions with 
that in ordinary plebeian inscriptions will reach the conclusion that 
even in these households the slave doorkeepers and cooks and hair- 
dressers and scullery-maids customarily married and had children. 
The volume is full of interesting instances : Livia's sarcinatrix mar- 
ried her mensor (VI. 3988), Octavia's ornatrix was the wife of her 
keeper of the plate (5539), Statilius's courier courted the spinning- 
maid of the household (6342). In the lists of husbands and wives 
one finds a chef (7458), a vestiarius (9963), a vestifica (5206), an 
unctor (6381), a slave-maid serving as secretary (a manu, 9540), 
the keeper of my lady's mirrors (7297), of her hand-bag (7368), of 
her wardrobe (4043), of her jewels (7296), and what not. Now, 
these inscriptions are all extremely brief. There are a great many 
like 4478, Domitia Sex. I. Artemisia, Tertius, Viator., where the 
word coniunx or contubernalis is probably, though not necessarily, 
understood. Furthermore, the record of children is not as complete 
as it would be in inscriptions of the better classes. A slave-child 
is, of course, not always honored with a record of its brief existence. 
Moreover, slave families, not being recognized in formal law, were 
sometimes broken up, so that some of the names fail to appear with 
the rest of the family. Nevertheless, the proportion of marriages 
and of offspring recorded by these very inscriptions, brief and in- 

Race Mixture in the Roman Empire 697 

complete as they are, is remarkably large. In the thousand inscrip- 
tions of the columbaria of the Livii, Drusi, Marcelli, and the first 
eighty of the Volusii (to make the even 1000) I find, 

151 inscriptions recording offspring. 

99 additional inscriptions recording marriage. 

152 additional inscriptions (like 4478 quoted above) probably 

recording marriage. 

Now this is not, of course, as large a proportion as is found in the 
main body of normal inscriptions. For comparison I give the pro- 
portions of 14,000 of volume VI., parts 2 and 3, reduced to the ratio 
of 1000: 

Per 1000 Total 

280 3923 inscriptions recording offspring. 

184 2577 additional inscriptions recording marriage. 

39 548 additional inscriptions probably recording marriage. 


Here, as we should expect, the proportion of children is larger, and 
the long list of inscriptions bearing names of a man and a woman 
whose relationship is not defined yields in favor of a record of con- 
juges. But, as has been said, the slave inscriptions are far briefer 
and less complete than the others. 

To discover whether the lower proportion in the first list might 
be due to the brevity of the inscriptions, I compared it with the list 
of 460 inscriptions of greater length, edited in volume VI., part 2, 
8639 ff., as being ex familia Augusta. These inscriptions are longer, 
to be sure, because the persons designated had reached some degree 
of prosperity and could afford a few feet of sod with a separate 
stone. But even these slaves and freedmen were generally required 
to furnish close and persistent attention to their service. I have 
again given the numbers in the proportion of 1000 for the sake of 

Per 1000 Total 

2 9° !33 inscriptions recording offspring. 

220 101 additional inscriptions recording marriages. 

78 36 additional inscriptions probably recording marriages. 


From this list, if we may draw any conclusions from such small 
numbers, it would appear that the imperial slaves and freedmen were 
more productive than the ordinary citizens of Rome. And I see no 
reason for doubting that the proportions in the households of the 
Livii, Drusi, etc., would be nearly as large if the inscriptions were 

698 Tenney Frank 

full lapidary ones, instead of the short notices that were painted or 
cut upon the small space of an urn. 

Finally, for the sake of getting a fuller record regarding the 
poorer classes, I read 3000 inscriptions of the miscellaneous colum- 
baria that follow those of the aristocratic households. These are 
nos. 4881-7881 of volume VI., part 2. A very few of these inscrip- 
tions contain names of poor free-born citizens who associated with — 
in fact were probably related to — slaves and ex-slaves, but the pro- 
portion is so small that we may safely use this group for our present 
purpose. Three thousand inscriptions from miscellaneous colum- 

Per 1000 Total 

154 462 inscriptions recording offspring, 

ill 332 additional inscriptions recording marriage. 

73 220 additional inscriptions probably recording marriage. 

This group, consisting of the very briefest inscriptions, set up by the 
poorest of Rome's menial slaves, shows, as we might expect, the 
smallest birth and marriage rate. But when we compare it with that 
of the corresponding class engaged in the aristocratic and imperial 
households, the ratios fall only in proportion to the brevity and 
inadequacy of the record. 

To sum up, then, it would seem that not only were the slaves of 
the familia rustica permitted and encouraged to marry, as Varro and 
Columella indicate, but — what the literary sources fail to tell — that 
slaves and freedmen in the familia urbana did not differ from coun- 
try slaves in this respect. And, considering the poverty of those 
who raised these humble memorials, the brevity of the records, and 
the ease with which members of such families were separated, the 
ratio of offspring is strikingly large. We cannot be far from wrong 
if we infer that the slaves and freedmen 11 of the city were nearly 
as prolific as the free-born population. 

But however numerous the offspring of the servile classes, unless 
the Romans had been liberal in the practice of manumission, these 
people would not have merged with the civil population. Now, lit- 
erary and legal records present abundant evidence of an unusual 
liberality in this practice at Rome, and the facts need not be repeated 
after the full discussions of Wallon, Buckland, Friedlander, Dill, 

11 We cannot suppose that most of the children belong to the period subse- 
quent to the liberation of the parents. Very many of the liberti recorded were eman- 
cipated in old age, and throughout the empire manumission of slaves under 30 
years of age was discouraged (Buckland, Roman Law of Slavery, p. 542). In a 
large number of instances the form and contents of the inscriptions show that 
slave-fathers after emancipation paid the price for children and wife. 

Race Mixture in the Roman Empire 699 

Lemonnier, and Cicotti. If there were any doubt that the laws 
passed in the early empire for the partial restriction of manumission 
did not seriously check the practice, the statistics given at the begin- 
ning of the paper would allay it. When from eighty to ninety per 
cent, of the urban-born population proves to have been of servile 
extraction, we can only conclude that manumission was not seriously 
restricted. I may add that a count of all the slaves and freedmen 
in the familiae of the aristocratic households mentioned above 
showed that almost a half were liberti. It is difficult to believe that 
this proportion represents the usual practice, however, and, in fact, 
the figures must be used with caution. On the one hand, they may 
be too high, for many who served as slaves all their lives were manu- 
mitted only in old age, and it must also be recognized that slaves 
were less apt to be recorded than liberti. On the other hand, the 
figures may in some respects be too low, since there can be little 
doubt that the designation liberti was at times omitted on the simple 
urns, even though the subject had won his freedom. However, as 
far as the inscriptions furnish definite evidence, they tell the same 
tale as the writers of Rome, namely, that slaves were at all times 
emancipated in great numbers. 

When we consider whence these slaves came and of what stock 
they actually were, we may derive some aid from an essay by Bang, 
Die Herkunft der Romischen Sklaven. Bang has collected all the 
inscriptions like Damas, natione Syrus, and C. Ducenius C. lib. natus 
in Syria, which reveal the provenance of slaves. Of course, the 
number of inscriptions giving such information is relatively small, a 
few hundred in all. It should also be noticed that when a slave 
gives his nationality he shows a certain pride in it, which, in some 
cases at least, implies that he is not a normal slave of the mart, born 
in servitude, but rather a man of free birth who may have come into 
the trade by capture, abduction, or some other special way. How- 
ever, with this word of caution we may use Bang's statistics for what 
they are worth. 

A very large proportion in his list (seven-eighths of those dating 
in our era) came from within the boundaries of the empire. From 
this we may possibly infer that war-captives were comparatively rare 
during the empire, and that, though abduction and kidnapping sup- 
plied some of the trade, the large bulk of the slaves were actually 
reared from slave-parents. Doubtless slaves were reared with a 
view to profit in Greece and the Orient, as well as in Italy, and I see 
no reason for supposing that the situation there differed much from 
that of our Southern States where — for obvious economic reasons — 
the birth-rate of slaves was higher between 1800 and i860 than the 

yoo Tenney Frank 

birth-rate of their free descendants has been since then. An exami- 
nation of the names in Bang's list with reference to the provenance 
of the bearer will do something toward giving a criterion for judg- 
ing the source of Italian slaves not otherwise specified. In a very 
few cases a name appears which is not Greek or Latin but Semitic, 
Celtic, etc., according to the birthplace of the slave, as, for instance, 
Malchio, Zizas, Belatusa. Such names are rare and never cause any 
difficulty. Somewhat more numerous, and equally clear of inter- 
pretation, are the generic names that explicitly give the race of the 
bearer, like Syrus, Cappadox, Gallus, etc. In general, however, 
slaves have Greek or Latin names, and here difficulties arise, for it 
has by no means been certain whether or not these names had so- 
distinctively servile a connotation that they might be applied indis- 
criminately to captives from the North and West, as well as to the 
slaves of Italy and the East. Nevertheless, there seems to be a fairly 
uniform practice which differentiated between Greek and Latin 
names during the empire. Slaves from Greece, from Syria, from 
Asia Minor, including the province of Asia, Phrygia, Caria, Lycia, 
Pamphylia, Cappadocia, Bithynia, Paphlagonia, Galatia — that is, 
from regions where Greek was the language of commerce, regularly 
bore Greek, rather than Latin, names. Slaves from the North — 
from Germany to Dacia — as a rule bore Latin names. Presumably 
their own barbaric names were difficult to pronounce and Greek ones 
seemed inappropriate. Slaves from Spain and Gaul bore Latin and 
Greek names in about equal numbers. But here we must apparently 
discriminate. These provinces were old and commerce had brought 
into them many Oriental slaves from the market. It may be that 
the Greek names were applied mostly to slaves of Eastern extraction. 
This I should judge to be the case at least with the following: 
Ephesia (Bang, p. 239), Corinthus, Hyginus, Phoebus (his father's 
name is Greek), Eros (a Sevir Aug.), and Philocyrius (p. 240, 
Hiibner reads Philo, Cyprius). In general we may apply these cri- 
teria in trying in some measure to decide the provenance of slaves 
in Italy whose nativity is not specified : bearers of Greek names are 
in general from the East or descendants of Eastern slaves who have 
been in the West ; bearers of Latin names are partly captives of the 
North and West, partly, as we have seen from our Roman lists,. 
Easterners and descendants of Easterners who have received Latin 
names from their masters. 

Therefore, when the urban inscriptions show that seventy per 
cent, of the city slaves and freedmen bear Greek names and that a 
large proportion of the children who have Latin names have parents 
of Greek names, this at once implies that the East was the source of 

Race Mixture in the Roman Empire 701 

most of them, and with that inference Bang's conclusions entirely 
agree. In his list of slaves that specify their origin as being outside 
of Italy (during the empire), by far the larger portion came from 
the Orient, especially from Syria and the provinces of Asia Minor, 
with some from Egypt and Africa (which for racial classification 
may be taken with the Orient). Some are from Spain and Gaul, 
but a considerable proportion of these came originally from the East. 
Very few slaves are recorded from the Alpine and Danube provinces, 
while Germans rarely appear, except among the imperial bodyguard. 
Bang remarks that Europeans were of greater service to the empire 
as soldiers than as servants. This is largely true, but, as Strack has 
commented, 12 the more robust European war-captives were apt to be 
chosen for the gruelling work in the mines and in industry, and con- 
sequently they have largely vanished from the records. Such slaves 
were probably also the least productive of the class ; and this, in turn, 
helps to explain the strikingly Oriental aspect of the new population. 

Up to this point we have dealt mainly with the inscriptions of 
the city. But they, of course, do not represent the state of affairs 
in the empire at large. Unfortunately, it is difficult to secure large 
enough groups of sepulchral inscriptions for other cities and districts 
to yield reliable average on the points just discussed. However, 
since the urban inscriptions have presented a general point of view 
regarding the prolificness of slaves and the significance of the Greek 
cognomen, it will suffice to record the proportion of servile and 
Oriental names found in some typical district outside of the city. 
The proportion of Greek names to Latin among the slaves and liberti 
of the city was, in the inscriptions I recorded, seventy per cent, 
versus thirty per cent. This is of course very high. In CIL., vol- 
ume XIV. (Latium outside of Rome), the index of cognomina gives 
571 to 315, that is, about sixty-four per cent, to thirty-six per cent. ; 
volume IX. (Calabria to Picenum), 810 to 714, i.e., fifty-three to 
forty-seven per cent.; volume V. (Cisalpine Gaul), 701 to 831, i.e., 
forty-six to fifty-four per cent. This, in fact, is the only part of 
Italy where the majority of slaves and freedmen recorded did not 
bear Greek names. As is to be expected, northern slaves, who gen- 
erally received Latin names, were probably found in larger numbers 
here ; but again it should not be forgotten that a great many of the 
Latin-named slaves were of Eastern extraction. 

In order to get more specific evidence regarding the nature of the 
population in the West, free as well as servile, we may read the 
sepulchral inscriptions of some typical towns 13 and districts. I have 

1 2 Historische Zeitschrift, CXII. 9. 

13 In this list I have omitted imperial officials and soldiers, since they are 
not likely to be natives of the place. 

702 Tenney Frank 

listed them in four groups : ( i ) slaves and f reedmen bearing Latin 
names; (2) slaves and f reedmen bearing Greek names; (3) free- 
born citizens with Latin cognomen; (4) free-born citizens with 
Greek cognomen. Under 3 and 4, I have, except when explicit evi- 
dence proved the contrary, credited the tria nomina as indication of 
free birth, but wish again to call attention to the caution contained 
in note 3. In cases of doubt the absence of the gentile name has 
been taken as an indication of servile station if the name given is 
Greek or Latin and not Barbarian. 

13 3 4 Sum 

Marsi and Vestini, Italy 201 119 234 58 612 

Beneventum, Italy 141 129 297 57 624 

Milan and Patavium, North Italy. . 182 135 400 93 810 

Narbo, Gaul 257 160 332 95 844 

Gades, Corduba 1 <, _ <■ 

Hispalis, Emerita j S P aln ^ ™ j™ _* _^5 

910 644 1568 393 3515 

When the indexes of CIL. are nearer completion such details will 
be more readily available and the tedious work of getting full sta- 
tistics may be undertaken with the hope of reaching some degree of 
finality. However, the trend is evident in what we have given, 
and the figures are, I think, fairly representative of the whole. In 
these towns, as at Rome, the proportion of non-Latin folk is strik- 
ingly large. Slaves, f reedmen, and citizens of Greek name make up 
more than half the population, despite the fact that in the nature of 
the case these are presumably the people least likely to be adequately 
represented in inscriptions. Furthermore, if the Latin names of 
freedmen in half the instances conceal persons of Oriental parent- 
age, as they do in the city, the Easterner would be represented by 
classes 2 and 4, half of class 1, and a part of class 3. How strik- 
ingly un-Latin these places must have appeared to those who saw 
the great crowd of humble slaves, who were buried without cere- 
mony or record in nameless trenches! Yet here are the Marsi, 
proverbially the hardiest native stock of the Italian mountains; 
Beneventum, one of Rome's old frontier colonies ; Milan and Padua, 
that drew Latins and Romanized Celts from the richest agricultural 
districts of the Po valley; the old colony of Narbo, the home of 
Caesar's famous Tenth Legion — the city that Cicero called specula 
populi Romani; and four cities at the western end of the empire. 
If we may, as I think fair, infer for these towns what we found to 
be true at Rome, namely, that slaves were quite as prolific as the 
civil population, that they merged into the latter, and that Greek 
names betokened Oriental stock, it is evident that the whole empire 

Race Mixture in the Roman Empire 703 

was a melting-pot and that the Oriental was always and everywhere 
a very large part of the ore. 

There are other questions that enter into the problem of change 
of race at Rome, for the solution of which it is even more difficult to 
obtain statistics. For instance, one asks, without hope of a sufficient 
answer, why the native stock did not better hold its own. Yet there 
are at hand not a few reasons. We know for instance that when 
Italy had been devastated by Hannibal and a large part of its popu- 
lation put to the sword, immense bodies of slaves were bought up in 
the East to fill the void; and that during the second century, when 
the plantation system with its slave service was coming into vogue, 
the natives were pushed out of the small farms and many disap- 
peared to the provinces of the ever-expanding empire. Thus, 
during the thirty years before Tiberius Gracchus, the census statis- 
tics show no increase. During the first century B.C., the importa- 
tion of captives and slaves continued, while the free-born citizens 
were being wasted in the social, Sullan, and civil wars. Augustus 
affirms that he had had half a million citizens under arms, one-eighth 
of Rome's citizens, and that the most vigorous part. During the early 
empire, twenty to thirty legions, drawn of course from the best free 
stock, spent their twenty years of vigor in garrison duty, while the 
slaves, exempt from such services, lived at home and increased in 
number. In other words, the native stock was supported by less 
than a normal birth-rate, whereas the stock of foreign extraction 
had not only a fairly normal birth-rate but a liberal quota of manu- 
missions to its advantage. Various other factors, more difficult to 
estimate, enter into the problem of the gradual attrition of the native 
stock. It seems clear, for instance, that the old Indo-Germanic 
custom of " exposing " children never quite disappeared from Rome. 
Law early restrained the practice and in the empire it was not per- 
mitted to expose normal males, and at least the first female must be 
reared. It is impossible, however, to form any clear judgment from 
the literary sources as to the extent of this practice during the em- 
pire. I thought that a count of the offspring in a large number of 
inscriptions might throw light upon the question, and found that of 
the 5063 children noted in the 19,000 inscriptions read, 3155, or 
about 62.3 per cent., were males. Perhaps this reflects the operation 
of the law in question, and shows that the expositio of females was 
actually practised to some extent. But here too we must remember 
that the evidence is, by its very nature, of little worth. Boys natu- 
rally had a better chance than girls to gain some little distinction and 
were therefore more apt to leave a sepulchral record. At any rate, 
if expositio was practised, the inscriptions show little difference in 

704 Tenney Frank 

this respect between the children of slaves and freedmen and the 
children of the ordinary city populace. 14 

But the existence of other forms of " race suicide ", so freely 
gossipped about by writers of the empire, also enters into this 
question, and here the inscriptions quite fail us. The importance of 
this consideration must, nevertheless, be kept in mind. Doubtless, 
as Fustel de Coulanges {La Cite Antique) has remarked, it could 
have been of little importance in the society of the republic so long 
as the old orthodox faith in ancestral spirits survived, for the happi- 
ness of the manes depended upon the survival of the family, and this 
religious incentive probably played the same role in the propagation 
of the race as the Mosaic injunctions among the Hebrews, which so 
impressed Tacitus in a more degenerate day of Rome. But reli- 
gious considerations and customs — which in this matter emanate 
from the fundamental instincts that continue the race — were ques- 
tioned as all else was questioned before Augustus's day. Then the 
process of diminution began. The significance of this whole ques- 
tion lies in the fact that " race suicide " then, as now, curtailed the 
stock of the more sophisticated, that is, of the aristocracy and the 
rich, who were, to a large extent, the native stock. Juvenal, satirist 
though he is, may be giving a fact of some social importance when 
he writes that the poor bore all the burdens of family life, while the 
rich remained childless : 

jacet aurato vix ulla puerpera lecto; 
Tantum artes hujus, tantum medicamina possunt, 
Quae steriles facit. 15 

There may lie here — rare phenomenon — an historic parallel of 
some meaning. The race of the human animal survives by means 
of instincts that shaped themselves for that purpose long before 
rational control came into play. Before our day it has only been at 
Greece and Rome that these impulses have had to face the obstacle 
of sophistication. There at least the instinct was beaten, and the 
race went under. The legislation of Augustus and his successors, 
while aimed at preserving the native stock, was of the myopic kind 
so usual in social law-making, and, failing to reckon with the real 
nature of the problem involved, it utterly missed the mark. By 
combining epigraphical and literary references, a fairly full history 
of the noble families can be procured, and this reveals a startling 
inability of such families to perpetuate themselves. We know, for 

« I have compared the respective ratios of the girls and boys of the Julii 
and the Claudii with those of the Aelii and the Aurelii (who would in general 
date about a century later) but found no appreciable difference in the percentage. 
A chronological test seems to be unattainable. 

15 VI. 594-596. 

Race Mixture in the Roman Empire 705 

instance, in Caesar's day of forty-five patricians, only one of whom 
is represented by posterity when Hadrian came to power. 16 The 
Aemilii, Fabii, Claudii, Manlii, Valerii, and all the rest, with the 
exception of the Cornelii, have disappeared. Augustus and Claudius 
raised twenty-five families to the patriciate, and all but six of them 
disappear before Nerva's reign. Of the families of nearly four 
hundred senators recorded in 65 A.D. under Nero, all trace of a half 
is lost by Nerva's day, a generation later. And the records are so 
full that these statistics may be assumed to represent with a fair 
degree of accuracy the disappearance of the male stock of the fami- 
lies in question. Of course members of the aristocracy were the 
chief sufferers from the tyranny of the first century, but this havoc 
was not all wrought by delatores and assassins. The voluntary 
choice of childlessness accounts largely for the unparalleled con- 
dition. This is as far as the records help upon this problem, which, 
despite the silence, is probably the most important phase of the 
whole question of the change of race. Be the causes what they 
may, the rapid decrease of the old aristocracy and the native stock 
was clearly concomitant with a twofold increase from below: by a 
more normal birth-rate of the poor, and the constant manumission 
of slaves. 

This Orientalizing of Rome's populace has a more important 
bearing than is usually accorded it upon the larger question of why 
the spirit and acts of imperial Rome are totally different from those 
of the republic, if indeed racial characteristics are not wholly a 
myth. There is to-day a healthy activity in the study of the eco- 
nomic factors — unscientific finance, fiscal agriculture, inadequate 
support of industry and commerce, etc. — that contributed to Rome's 
decline. But what lay behind and constantly reacted upon all such 
causes of Rome's disintegration was, after all, to a considerable ex- 
tent, the fact that the people who built Rome had given way to a 
different race. The lack of energy and enterprise, the failure of 
foresight and common sense, the weakening of moral and political 
stamina, all were concomitant with the gradual diminution of the 
stock which, during the earlier days, had displayed these qualities. 
It would be wholly unfair to pass judgment upon the native qualities 
of the Orientals without a further study, or to accept the self- 
complacent slurs of the Romans, who, ignoring certain imaginative 
and artistic qualities, chose only to see in them unprincipled and 
servile egoists. We may even admit that had the new races had 
time to amalgamate and attain a political consciousness, a more bril- 
liant and versatile civilization might have come to birth. That, 

16 Stech, in Klio, Beiheft X. 
AM. HIST. REV., VOL. XXI. — 46. 

J06 Tenney Frank 

however, is not the question. It is apparent that at least the political 
and moral qualities which counted most in the building of the Italian 
federation, the army organization, the provincial administrative sys- 
tem of the republic, were the qualities most needed in holding the 
empire together. And however brilliant the endowment of the new 
citizens, these qualities they lacked. The Trimalchios of the empire 
were often shrewd and daring business men, but their first and ob- 
vious task apparently was to climb by the ladder of quick profits to 
a social position in which their children with Romanized names could 
comfortably proceed to forget their forebears. The possession of 
wealth did not, as in the republic, suggest certain duties toward the 
commonwealth. Narcissus and Pallas might be sagacious politi- 
cians, but they were not expected to be statesmen concerned with 
the continuity of the mos ma jorum. And when, on reading Tacitus, 
we are amazed at the new servility of Scipios and Messalas, we must 
recall that these scattered inheritors of the old aristocratic ideals 
had at their back only an alien rabble of ex-slaves, to whom they 
would have appealed in vain for a return to ancestral ideas of law 
and order. They had little choice between servility and suicide, and 
not a few chose the latter. 

It would be illuminating by way of illustration of this change to 
study the spread of the mystery religions. Cumont seems to think 
that these cults won many converts among all classes in the West. 
Toutain, skeptical on this point, assigns not a little of the new reli- 
gious activity to the rather formal influence of the court at Rome. 
Dobschutz, a more orthodox churchman, seems to see in the spread 
of these cults the pervasion of a new and deeper religious spirit, 
which, in some mystical way, was preparing the old world for Chris- 
tianity. But is not the success of the cults in great measure an 
expression of the religious feelings of the new people themselves? 
And if it is, may it not be that Occidentals who are actually of 
Oriental extraction, men of more emotional nature, are simply find- 
ing in these cults the satisfaction that, after long deprivation, their 
temperaments naturally required? When a senator, dignified by 
the name of M. Aurelius Victor, is found among the votaries of 
Mithras in the later empire, it may well be that he is the great- 
grandson of some child kidnapped in Parthia and sold on the block 
at Rome. Toutain has proved, I think, that in the northern and 
western provinces the only Oriental cult that took root at all among 
the real natives was that of Magna Mater, and this goddess, whose 
cult was directed by the urban priestly board, had had the advantage 
of centuries of a rather accidental recognition by the Roman state. 
In the western provinces, the Syrian and Egyptian gods were wor- 

Race Mixture in the Roman Empire joy 

shipped chiefly by people who seem not to be native to the soil. 
The Mithraic worshippers in these provinces were, for the most part, 
soldiers recruited or formerly stationed in the East, and Orientals 
who, by way of commerce or the slave-market, had come to live in 
the West. From the centres where such people lived the cult 
spread but very slowly. 

It would hardly be worth while to attempt any conclusion for 
the city of Rome, since, as we have seen, the whole stock there had 
so changed that fair comparisons would be well-nigh unattainable; 
but the Po valley, that is Cisalpine Gaul, which preserved its Occi- 
dental aspect better than any other part of Italy, might yield usable 
data. For this region nearly one hundred devotees of Oriental gods 
are recorded in the fifth volume of CIL., and, as soldiers and Roman 
officers are not numerous there, the worshippers may be assumed to 
represent a normal average for the community. Among them I find 
only twelve who are actually recorded as slaves or freedmen, but 
upon examination of the names, more than four-fifths seem, after 
all, to belong to foreign stock. Nearly half have Greek names. 
Several are seviri Augustales, and, therefore, probably liberti; and 
names like Publicius, Verna, Veronius (at Verona), tell the same 
tale. Finally, there are several imperial gentile names — Claudius, 
Flavius, Ulpius, Aelius, etc. — which, when found among such people, 
suggest that the Roman nomenclature is a recent acquisition. There 
is a residue of only some twelve names the antecedents of which re- 
main undefined. This seems to me to be a fairly typical situation, 
and not without significance. In short, the mystery cults permeated 
the city, Italy, and the western provinces only to such an extent as 
the city and Italy and the provinces were permeated by the stock 
that had created those religions. 

At Rome, Magna Mater was introduced for political reasons 
during the Punic War, when the city was still Italian. The rites 
proved to be shocking to the unemotional westerner, who worshipped 
the staid patrician called Jupiter Optimus Maximus, and were locked 
in behind a wall. As the urban populace began to change, however, 
new rites clamored for admittance, for, as a senator in Nero's days 
says, 17 " Nationes in familiis habemus, quibus diversi ritus, externa 
sacra." And as the populace enforced their demands upon the em- 
peror for panem et circenses, so they also secured recognition for 
their externa sacra. One after another of the emperors gained 
popularity with the rabble by erecting a shrine to some foreign Baal, 
or a statue to Isis in his chapel, in much the same way that our cities 
are lining their park drives with tributes to Garibaldi, Pulaski, and 

i' Tacitus, Annates, XIV. 44. 

708 Tenney Frank 

who knows what -vitch. Finally, in the third and fourth centuries, 
when even the aristocracy at Rome was almost completely foreign, 
these Eastern cults, rather than those of old Rome, became the cen- 
tres of " patrician " opposition to Christianity. In other words, the 
western invasion of the mystery cults is hardly a miraculous conver- 
sion of the even-tempered, practical-minded Indo-European to an 
orgiastic emotionalism, foreign to his nature. These religions came 
with their peoples, and in so far as they gained new converts, they 
attracted for the most part people of Oriental extraction who had 
temporarily fallen away from native ways in the western world. 
Christianity, which contained enough Oriental mysticism to appeal 
to the vast herd of Easterners in the West, and enough Hellenic 
sanity to captivate the rationalistic Westerner, found, even if one 
reckons only with social forces, the most congenial soil for growth 
in the conglomeration of Europeans, Asiatics, and Africans that 
filled the western Roman Empire in the second century. 

This is but one illustration. But it is offered in the hope that a 
more thorough study of the race question may be made in conjunc- 
tion with economic and political questions before any attempt is 
made finally to estimate the factors at work in the change of temper 
of imperial Rome. 

Tenney Frank.